Reference

Edit This Page

kubeadm init

This command initializes a Kubernetes control-plane node.

Run this command in order to set up the Kubernetes master.

Synopsis

Run this command in order to set up the Kubernetes master.

The “init” command executes the following phases:

preflight                  Run master pre-flight checks
kubelet-start              Writes kubelet settings and (re)starts the kubelet
certs                      Certificate generation
  /ca                        Generates the self-signed Kubernetes CA to provision identities for other Kubernetes components
  /apiserver                 Generates the certificate for serving the Kubernetes API
  /apiserver-kubelet-client  Generates the Client certificate for the API server to connect to kubelet
  /front-proxy-ca            Generates the self-signed CA to provision identities for front proxy
  /front-proxy-client        Generates the client for the front proxy
  /etcd-ca                   Generates the self-signed CA to provision identities for etcd
  /etcd-server               Generates the certificate for serving etcd
  /etcd-peer                 Generates the credentials for etcd nodes to communicate with each other
  /etcd-healthcheck-client   Generates the client certificate for liveness probes to healtcheck etcd
  /apiserver-etcd-client     Generates the client apiserver uses to access etcd
  /sa                        Generates a private key for signing service account tokens along with its public key
kubeconfig                 Generates all kubeconfig files necessary to establish the control plane and the admin kubeconfig file
  /admin                     Generates a kubeconfig file for the admin to use and for kubeadm itself
  /kubelet                   Generates a kubeconfig file for the kubelet to use *only* for cluster bootstrapping purposes
  /controller-manager        Generates a kubeconfig file for the controller manager to use
  /scheduler                 Generates a kubeconfig file for the scheduler to use
control-plane              Generates all static Pod manifest files necessary to establish the control plane
  /apiserver                 Generates the kube-apiserver static Pod manifest
  /controller-manager        Generates the kube-controller-manager static Pod manifest
  /scheduler                 Generates the kube-scheduler static Pod manifest
etcd                       Generates static Pod manifest file for local etcd.
  /local                     Generates the static Pod manifest file for a local, single-node local etcd instance.
upload-config              Uploads the kubeadm and kubelet configuration to a ConfigMap
  /kubeadm                   Uploads the kubeadm ClusterConfiguration to a ConfigMap
  /kubelet                   Uploads the kubelet component config to a ConfigMap
mark-control-plane         Mark a node as a control-plane
bootstrap-token            Generates bootstrap tokens used to join a node to a cluster
addon                      Installs required addons for passing Conformance tests
  /coredns                   Installs the CoreDNS addon to a Kubernetes cluster
  /kube-proxy                Installs the kube-proxy addon to a Kubernetes cluster
kubeadm init [flags]

Options

--apiserver-advertise-address string
The IP address the API Server will advertise it's listening on. Specify '0.0.0.0' to use the address of the default network interface.
--apiserver-bind-port int32     Default: 6443
Port for the API Server to bind to.
--apiserver-cert-extra-sans stringSlice
Optional extra Subject Alternative Names (SANs) to use for the API Server serving certificate. Can be both IP addresses and DNS names.
--cert-dir string     Default: "/etc/kubernetes/pki"
The path where to save and store the certificates.
--config string
Path to kubeadm config file. WARNING: Usage of a configuration file is experimental.
--cri-socket string     Default: "/var/run/dockershim.sock"
Specify the CRI socket to connect to.
--dry-run
Don't apply any changes; just output what would be done.
--feature-gates string
A set of key=value pairs that describe feature gates for various features. Options are:
-h, --help
help for init
--ignore-preflight-errors stringSlice
A list of checks whose errors will be shown as warnings. Example: 'IsPrivilegedUser,Swap'. Value 'all' ignores errors from all checks.
--image-repository string     Default: "k8s.gcr.io"
Choose a container registry to pull control plane images from
--kubernetes-version string     Default: "stable-1"
Choose a specific Kubernetes version for the control plane.
--node-name string
Specify the node name.
--pod-network-cidr string
Specify range of IP addresses for the pod network. If set, the control plane will automatically allocate CIDRs for every node.
--service-cidr string     Default: "10.96.0.0/12"
Use alternative range of IP address for service VIPs.
--service-dns-domain string     Default: "cluster.local"
Use alternative domain for services, e.g. "myorg.internal".
--skip-phases stringSlice
List of phases to be skipped
--skip-token-print
Skip printing of the default bootstrap token generated by 'kubeadm init'.
--token string
The token to use for establishing bidirectional trust between nodes and masters. The format is [a-z0-9]{6}\.[a-z0-9]{16} - e.g. abcdef.0123456789abcdef
--token-ttl duration     Default: 24h0m0s
The duration before the token is automatically deleted (e.g. 1s, 2m, 3h). If set to '0', the token will never expire

Options inherited from parent commands

--rootfs string
[EXPERIMENTAL] The path to the 'real' host root filesystem.

Init workflow

kubeadm init bootstraps a Kubernetes control-plane node by executing the following steps:

  1. Runs a series of pre-flight checks to validate the system state before making changes. Some checks only trigger warnings, others are considered errors and will exit kubeadm until the problem is corrected or the user specifies --ignore-preflight-errors=<list-of-errors>.

  2. Generates a self-signed CA (or using an existing one if provided) to set up identities for each component in the cluster. If the user has provided their own CA cert and/or key by dropping it in the cert directory configured via --cert-dir (/etc/kubernetes/pki by default) this step is skipped as described in the Using custom certificates document. The APIServer certs will have additional SAN entries for any --apiserver-cert-extra-sans arguments, lowercased if necessary.

  3. Writes kubeconfig files in /etc/kubernetes/ for the kubelet, the controller-manager and the scheduler to use to connect to the API server, each with its own identity, as well as an additional kubeconfig file for administration named admin.conf.

  4. Generates static Pod manifests for the API server, controller manager and scheduler. In case an external etcd is not provided, an additional static Pod manifest are generated for etcd.

Static Pod manifests are written to /etc/kubernetes/manifests; the kubelet watches this directory for Pods to create on startup.

Once control plane Pods are up and running, the kubeadm init sequence can continue.

  1. Apply labels and taints to the control-plane node so that no additional workloads will run there.

  2. Generates the token that additional nodes can use to register themselves with the master in the future. Optionally, the user can provide a token via --token, as described in the kubeadm token docs.

  3. Makes all the necessary configurations for allowing node joining with the Bootstrap Tokens and TLS Bootstrap mechanism:

    • Write a ConfigMap for making available all the information required for joining, and set up related RBAC access rules.

    • Let Bootstrap Tokens access the CSR signing API.

    • Configure auto-approval for new CSR requests.

See kubeadm join for additional info.

  1. Installs a DNS server (CoreDNS) and the kube-proxy addon components via the API server. In Kubernetes version 1.11 and later CoreDNS is the default DNS server. To install kube-dns instead of CoreDNS, the DNS addon has to configured in the kubeadm ClusterConfiguration. For more information about the configuration see the section Using kubeadm init with a configuration file bellow. Please note that although the DNS server is deployed, it will not be scheduled until CNI is installed.

Using init phases with kubeadm

Kubeadm allows you create a control-plane node in phases. In 1.13 the kubeadm init phase command has graduated to GA from it’s previous alpha state under kubeadm alpha phase.

To view the ordered list of phases and sub-phases you can call kubeadm init --help. The list will be located at the top of the help screen and each phase will have a description next to it. Note that by calling kubeadm init all of the phases and sub-phases will be executed in this exact order.

Some phases have unique flags, so if you want to have a look at the list of available options add --help, for example:

sudo kubeadm init phase control-plane controller-manager --help

You can also use --help to see the list of sub-phases for a certain parent phase:

sudo kubeadm init phase control-plane --help

kubeadm init also expose a flag called --skip-phases that can be used to skip certain phases. The flag accepts a list of phase names and the names can be taken from the above ordered list.

An example:

sudo kubeadm init phase control-plane all --config=configfile.yaml
sudo kubeadm init phase etcd local --config=configfile.yaml
# you can now modify the control plane and etcd manifest files
sudo kubeadm init --skip-phases=control-plane,etcd --config=configfile.yaml

What this example would do is write the manifest files for the control plane and etcd in /etc/kubernetes/manifests based on the configuration in configfile.yaml. This allows you to modify the files and then skip these phases using --skip-phases. By calling the last command you will create a control plane node with the custom manifest files.

Using kubeadm init with a configuration file

Caution: Caution: The config file is still considered beta and may change in future versions.

It’s possible to configure kubeadm init with a configuration file instead of command line flags, and some more advanced features may only be available as configuration file options. This file is passed in the --config option.

In Kubernetes 1.11 and later, the default configuration can be printed out using the kubeadm config print command. It is recommended that you migrate your old v1alpha3 configuration to v1beta1 using the kubeadm config migrate command, because v1alpha3 will be removed in Kubernetes 1.14.

For more details on each field in the v1beta1 configuration you can navigate to our API reference pages.

Adding kube-proxy parameters

For information about kube-proxy parameters in the kubeadm configuration see: - kube-proxy

For information about enabling IPVS mode with kubeadm see: - IPVS

Passing custom flags to control plane components

For information about passing flags to control plane components see: - control-plane-flags

Using custom images

By default, kubeadm pulls images from k8s.gcr.io, unless the requested Kubernetes version is a CI version. In this case, gcr.io/kubernetes-ci-images is used.

You can override this behavior by using kubeadm with a configuration file. Allowed customization are:

Please note that the configuration field kubernetesVersion or the command line flag --kubernetes-version affect the version of the images.

Using custom certificates

By default, kubeadm generates all the certificates needed for a cluster to run. You can override this behavior by providing your own certificates.

To do so, you must place them in whatever directory is specified by the --cert-dir flag or CertificatesDir configuration file key. By default this is /etc/kubernetes/pki.

If a given certificate and private key pair exists, kubeadm skips the generation step and existing files are used for the prescribed use case. This means you can, for example, copy an existing CA into /etc/kubernetes/pki/ca.crt and /etc/kubernetes/pki/ca.key, and kubeadm will use this CA for signing the rest of the certs.

External CA mode

It is also possible to provide just the ca.crt file and not the ca.key file (this is only available for the root CA file, not other cert pairs). If all other certificates and kubeconfig files are in place, kubeadm recognizes this condition and activates the “External CA” mode. kubeadm will proceed without the CA key on disk.

Instead, run the controller-manager standalone with --controllers=csrsigner and point to the CA certificate and key.

Managing the kubeadm drop-in file for the kubelet

The kubeadm package ships with configuration for how the kubelet should be run. Note that the kubeadm CLI command never touches this drop-in file. This drop-in file belongs to the kubeadm deb/rpm package.

This is what it looks like:

[Service]
Environment="KUBELET_KUBECONFIG_ARGS=--bootstrap-kubeconfig=/etc/kubernetes/bootstrap-kubelet.conf --kubeconfig=/etc/kubernetes/kubelet.conf"
Environment="KUBELET_SYSTEM_PODS_ARGS=--pod-manifest-path=/etc/kubernetes/manifests --allow-privileged=true"
Environment="KUBELET_NETWORK_ARGS=--network-plugin=cni --cni-conf-dir=/etc/cni/net.d --cni-bin-dir=/opt/cni/bin"
Environment="KUBELET_DNS_ARGS=--cluster-dns=10.96.0.10 --cluster-domain=cluster.local"
Environment="KUBELET_AUTHZ_ARGS=--authorization-mode=Webhook --client-ca-file=/etc/kubernetes/pki/ca.crt"
Environment="KUBELET_CADVISOR_ARGS="
Environment="KUBELET_CERTIFICATE_ARGS=--rotate-certificates=true --cert-dir=/var/lib/kubelet/pki"
ExecStart=/usr/bin/kubelet $KUBELET_KUBECONFIG_ARGS $KUBELET_SYSTEM_PODS_ARGS $KUBELET_NETWORK_ARGS $KUBELET_DNS_ARGS $KUBELET_AUTHZ_ARGS $KUBELET_CADVISOR_ARGS $KUBELET_CERTIFICATE_ARGS $KUBELET_EXTRA_ARGS

Here’s a breakdown of what/why:

Use kubeadm with CRI runtimes

Since v1.6.0, Kubernetes has enabled the use of CRI, Container Runtime Interface, by default. The container runtime used by default is Docker, which is enabled through the built-in dockershim CRI implementation inside of the kubelet.

Other CRI-based runtimes include:

Refer to the CRI installation instructions for more information.

After you have successfully installed kubeadm and kubelet, execute these two additional steps:

  1. Install the runtime shim on every node, following the installation document in the runtime shim project listing above.

  2. Configure kubelet to use the remote CRI runtime. Please remember to change RUNTIME_ENDPOINT to your own value like /var/run/{your_runtime}.sock:

cat > /etc/systemd/system/kubelet.service.d/20-cri.conf <<EOF
[Service]
Environment="KUBELET_EXTRA_ARGS=--container-runtime=remote --container-runtime-endpoint=$RUNTIME_ENDPOINT"
EOF
systemctl daemon-reload

Now kubelet is ready to use the specified CRI runtime, and you can continue with the kubeadm init and kubeadm join workflow to deploy Kubernetes cluster.

You may also want to set --cri-socket to kubeadm init and kubeadm reset when using an external CRI implementation.

Using internal IPs in your cluster

In order to set up a cluster where the master and worker nodes communicate with internal IP addresses (instead of public ones), execute following steps.

  1. When running init, you must make sure you specify an internal IP for the API server’s bind address, like so:

kubeadm init --apiserver-advertise-address=<private-master-ip>

  1. When a master or worker node has been provisioned, add a flag to /etc/systemd/system/kubelet.service.d/10-kubeadm.conf that specifies the private IP of the worker node:

--node-ip=<private-node-ip>

  1. Finally, when you run kubeadm join, make sure you provide the private IP of the API server addressed as defined in step 1.

Setting the node name

By default, kubeadm assigns a node name based on a machine’s host address. You can override this setting with the --node-nameflag. The flag passes the appropriate --hostname-override to the kubelet.

Be aware that overriding the hostname can interfere with cloud providers.

Self-hosting the Kubernetes control plane

As of 1.8, you can experimentally create a self-hosted Kubernetes control plane. This means that key components such as the API server, controller manager, and scheduler run as DaemonSet pods configured via the Kubernetes API instead of static pods configured in the kubelet via static files.

To create a self-hosted cluster, pass the flag --feature-gates=SelfHosting=true to kubeadm init.

Caution: SelfHosting is an alpha feature. It is deprecated in 1.12 and will be removed in 1.13.

Caveats

Self-hosting in 1.8 and later has some important limitations. In particular, a self-hosted cluster cannot recover from a reboot of the control-plane node without manual intervention. This and other limitations are expected to be resolved before self-hosting graduates from alpha.

By default, self-hosted control plane Pods rely on credentials loaded from hostPath volumes. Except for initial creation, these credentials are not managed by kubeadm.

In kubeadm 1.8, the self-hosted portion of the control plane does not include etcd, which still runs as a static Pod.

Process

The self-hosting bootstrap process is documented in the kubeadm design document.

In summary, kubeadm alpha selfhosting works as follows:

  1. Waits for this bootstrap static control plane to be running and healthy. This is identical to the kubeadm init process without self-hosting.

  2. Uses the static control plane Pod manifests to construct a set of DaemonSet manifests that will run the self-hosted control plane. It also modifies these manifests where necessary, for example adding new volumes for secrets.

  3. Creates DaemonSets in the kube-system namespace and waits for the resulting Pods to be running.

  4. Once self-hosted Pods are operational, their associated static Pods are deleted and kubeadm moves on to install the next component. This triggers kubelet to stop those static Pods.

  5. When the original static control plane stops, the new self-hosted control plane is able to bind to listening ports and become active.

Running kubeadm without an internet connection

For running kubeadm without an internet connection you have to pre-pull the required master images for the version of choice:

Image Name v1.10 release branch version
k8s.gcr.io/kube-apiserver-${ARCH} v1.10.x
k8s.gcr.io/kube-controller-manager-${ARCH} v1.10.x
k8s.gcr.io/kube-scheduler-${ARCH} v1.10.x
k8s.gcr.io/kube-proxy-${ARCH} v1.10.x
k8s.gcr.io/etcd-${ARCH} 3.1.12
k8s.gcr.io/pause-${ARCH} 3.1
k8s.gcr.io/k8s-dns-sidecar-${ARCH} 1.14.8
k8s.gcr.io/k8s-dns-kube-dns-${ARCH} 1.14.8
k8s.gcr.io/k8s-dns-dnsmasq-nanny-${ARCH} 1.14.8
coredns/coredns 1.0.6

Here v1.10.x means the “latest patch release of the v1.10 branch”.

${ARCH} can be one of: amd64, arm, arm64, ppc64le or s390x.

If you run Kubernetes version 1.10 or earlier, and if you set --feature-gates=CoreDNS=true, you must also use the coredns/coredns image, instead of the three k8s-dns-* images.

In Kubernetes 1.11 and later, you can list and pull the images using the kubeadm config images sub-command:

kubeadm config images list
kubeadm config images pull

Starting with Kubernetes 1.12, the k8s.gcr.io/kube-*, k8s.gcr.io/etcd and k8s.gcr.io/pause images don’t require an -${ARCH} suffix.

Automating kubeadm

Rather than copying the token you obtained from kubeadm init to each node, as in the basic kubeadm tutorial, you can parallelize the token distribution for easier automation. To implement this automation, you must know the IP address that the master will have after it is started.

  1. Generate a token. This token must have the form <6 character string>.<16 character string>. More formally, it must match the regex: [a-z0-9]{6}\.[a-z0-9]{16}.

    kubeadm can generate a token for you:

    kubeadm token generate
  2. Start both the control-plane node and the worker nodes concurrently with this token. As they come up they should find each other and form the cluster. The same --token argument can be used on both kubeadm init and kubeadm join.

Once the cluster is up, you can grab the admin credentials from the control-plane node at /etc/kubernetes/admin.conf and use that to talk to the cluster.

Note that this style of bootstrap has some relaxed security guarantees because it does not allow the root CA hash to be validated with --discovery-token-ca-cert-hash (since it’s not generated when the nodes are provisioned). For details, see the kubeadm join.

What's next

Feedback